Port-vax archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Issues with native build of -10 on VAX
On 2023-12-12 16.57, emu wrote:
On 2023-12-10 22:50, Greg Oster wrote:
Using a very recent -current source and your patch my simh test with
-current is up to libgmp, which is the furthest it's ever been in
recent memory! (I did a cross-build with a patched source tree, and
then installed the results on a simh instance. I am now doing a
'native' build with that simh instance. I also have a native build
happening on a physical 4000/60... but the results from the latter
will be weeks off..)
Thanks!
Later...
Greg Oster
Any chance you guys would share, how to set it up? I got a clean install
of NetBSD-10.0_RC1-vax on a 4000/90.
Now what?
Grab a copy of NetBSD-10.0_RC1 (or -current) sources (various options
here: https://www.netbsd.org/docs/guide/en/chap-fetch.html , e.g. from:
http://nycdn.netbsd.org/pub/NetBSD-daily/netbsd-10/202312120410Z/source/sets/
) and put them somewhere (they don't have to go into /usr/src ).
How do I get the gcc12 on it, with the required patches?
You don't need to get gcc12 -- You can just use the gcc that is with the
sources you fetch above. Grab the patch that Kalvis posted on Dec 7,
and patch the source.
Get is from PKGSRC first & install, than start a "build.sh -V HAVE_GCC12"?
Just use regular 'build.sh' - no need to specify a different compiler
once the compiler sources are patched... e.g.:
cd /u1/builds/build1/src/
sh ./build.sh -U -u -m vax -D /u1/builds/build1/install/vax \
-R /u1/builds/build1/release
if your sources are in /u1/builds/build1/src/. (My build line is actually:
sh ./build.sh -U -u -m vax -j 2 -V MAKECONF=/u1/builds/mk.conf \
-D /u1/builds/build3/install/vax -R /u1/builds/build3/release
as the build line gets generated and run by a different script...)
You'll want to have a healthy amount of swapspace setup... right now I'd
suggest 512MB or better. And set your shell limts to something like:
cputime unlimited
filesize unlimited
datasize 524288 kbytes
stacksize 8192 kbytes
coredumpsize 0 kbytes
memoryuse 58296 kbytes
memorylocked 19432 kbytes
maxproc 148
maxthread 2048
openfiles 548
sbsize unlimited
vmemoryuse 524288 kbytes
(which I do with this, in csh:
limit maxproc unlimited
limit datasize unlimited
limit stacksize unlimited
limit coredumpsize 0
limit openfiles unlimited
limit vmemoryuse unlimited
)
If you don't bump the default limits up building certain files will
fail... with '-j 2' I needed over 512MB to get the build to not fail -
1024MB seems better...
Sorry, my machines(and me?) collected a lot of dust in the last years ...
Hopefully the above helps clear the dust a bit :)
Later...
Greg Oster
Home |
Main Index |
Thread Index |
Old Index